DARE-bench (Data Science) - Classification (ML Modeling): leaderboard
Metric: Macro-F1 (shown times 100) against the ground-truth labels on the 74 open-ended classification modeling tasks; DARE-bench test tasks derived from recently updated Kaggle datasets; the model works as a data-science agent with a sandboxed Python execution tool (5 interaction turns, 200 s per execution, greedy decoding), mean of three repeats; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.7 Sonnet | 61.03 | #241 |
| 2 | O4 Mini | 57.89 | #172 |
| 3 | GPT-4.1 | 57.83 | #240 |
| 4 | GPT-5 | 43.4 | #91 |
| 5 | GPT-4o | 40.45 | #333 |
| 6 | Qwen 3 32B | 30.71 | #424 |
| 7 | Claude Sonnet 4 | 18.27 | #194 |
| 8 | Qwen 3 4B | 5.23 | #823 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=dare-bench-data-science-classification-ml-modeling · How It Works · Data refreshed daily, snapshot 2026-10-11.